employment and work task group full report

8/2/2019 Employment and Work Task Group Full Report

1/86

- 1 -

Employment arrangements, work conditions and health inequalities

Report on new evidence on health inequality reduction, produced by Task group 2 for the

Strategic review of health inequalities post 2010

Johannes Siegrist (chair)1, Joan Benach2, Abigail McKnight3 and Peter Goldblatt4

in collaboration with Carles Muntaner5

Research support

We are grateful to David MacFarlane, BSc, Barcelona; Monste Vergara Duarte, MPH,Barcelona; Hans Weitkowitz, Duesseldorf; and Gry Wester, MPhil, London for theirsupport in preparing this report.

1 Professor Johannes Siegrist, Ph.D., Director, Department of Medical Sociology,University of Duesseldorf, [email protected]

2 Professor Joan Benach, Ph.D., Health Inequalities Research Group, OccupationalHealth Research Center, Department of Experimental and Health Sciences, UniversityPompeu Fabra, Barcelona, Spain

[email protected] Abigail McKnight, Ph.D., CASE, London School of Economics, London, United

[email protected]

4 Peter Goldblatt, Ph.D., Department of Epidemiology and Public Health, UniversityCollege London, United [email protected]

5 Carles Muntaner, Ph.D., Dalla Lana School of Public Health, University of Toronto

and Institute for Work and Health, Toronto, [email protected]


2/86

- 2 -

Contents

Summary 3

1. Introduction 5

1.1. Employment, work and health: essential relations 5

1.2. A conceptual approach towards disentangling links between employment

relations and health inequalities7

1.3 The process of knowledge generation 8

1.4. Defining a health-adverse psychosocial environment 10

1.5. Structure of the report 12

2. The context for recommending future policies and interventions 132.1. Employment and working conditions in England 13

2.2. Recent initiatives on improving health and work in England 21

3. Answering the key questions 23

3.1. Health effects of adverse work and employment conditions: a selective review 23

3.2. The role of work and employment in explaining social inequalities in health 28

3.3. Health-promoting effects of work-related interventions 29

3.4. Socio-political and economic conditions and consequences of interventions 384. Recommendations and priorities 41

References 45

Appendix to Task Group 2 Report 63

Table 1: Overview of epidemiological studies: Job insecurity, downsizing and health

(1990-2008)64

Table 2: Overview of epidemiological studies: Temporary employment and health

(1990-2008)

69

Table 3: Overview of longitudinal observational studies: Associations of an adverse

psychosocial work environment (job strain, effort-reward imbalance) with physical

or mental disorder (in chronological order)

74

References to Tables 1, 2, 3 78


3/86

- 3 -

Summary

Employment and working conditions make a significant contribution to the development of

social inequalities in health in England, as is the case in all wealthy countries. They are of

critical importance to improve population health and redress health inequalities in several

interrelated ways. First, labour market and economic policies determine employment rates

and conditions (e.g. precarious, insecure or informal work). These have a major impact on a

range of life chances associated with paid work as a main social role in adult life. Second,

wages and salaries provide the main component of income. Low and insecure income affects

health via material deprivation, unhealthy behaviours and stressful experience. Importantly,

due to childhood antecedents of poor adult health, low income can have long-lasting negative

effects across generations. Third, adverse working conditions in terms of physical and

chemical hazards, risks of injuries, long or irregular work hours, shift work and physicallydemanding work affect workers health, defining targets of occupational health and safety

measures. Fourth, as the organisation of work and employment has changed significantly

during the last century, psychological and socio-emotional job demands and threats evolving

from insecure employment conditions and other forms of an adverse psychosocial work

environment have become more common. As these demands and threats have been shown to

directly affect the health of workers, new challenges have emerged to strengthen 'good'

(health-promoting and -protective) work through primary and secondary preventativemeasures.

The distribution of unemployment and health-adverse employment and working conditions

across the workforce is socially patterned, leaving those in lower socioeconomic positions at

higher risk. The scientific evidence for the associations between adverse employment and

working conditions and a range of indicators of poor health is summarised in this report, and

available results from intervention studies are presented and discussed. The implications for

policy- and workplace-related interventions are derived from this evidence.Finally, recommendations are proposed that are intended to make a significant contribution

towards healthier work and, ultimately, towards reducing social inequalities in health. These

recommendations are based on general principles that focus on fair employment and

improved quality and safety of work as a central goal of governmental policies. The reduction

of harmful employment and working conditions (through legislation, income transfers,

empowering workers, and integrating labour standards with labour market regulations) and

the implementation of participatory activities and inter-sectoral, contextualised interventions


4/86

- 4 -

at national, regional and local levels should be pursued in accordance with the principles of a

sustainable economy.

Specific recommendations concern measures to increase job security, to enforce protection in

employment, to enhance participation at work, to promote control and reward at work, to

reintegrate sick, disabled and unemployed people, and to strengthen the work-life balance.

These recommendations need to be further elaborated, harmonized with recommendations

from other task groups, evaluated with respect to their feasibility, measurement and

implementation, and prioritised in the context of short-term, medium-term, and long-term

policies. The medium and long-term perspectives should embody the principles of sustainable

development.


5/86

- 5 -

1. Introduction

1.1. Employment, work and health: essential relations

Work and employment make a significant contribution to the development of social

inequalities in health in England, as is the case in all modern societies. They are of critical

importance for population health and health inequalities in at least four interrelated ways:

First, participation in, or exclusion from the labour market determines a range of life chances

that are mainly mediated through regular wages and salaries. Adverse effects on health

produced by the exclusion from work and employment are most visible among those who

experience long-term unemployment (Bethune 1997). In addition to material constraints and

deprivations resulting from loss of employment, many psychosocial stressors contribute to

poor health not only among the unemployed themselves, but also among their partners and

children (Bartley et al 2006). These constraints and stressors are related to the loss of a core

role in social life that is crucial for ones sense of identity, thus prevention goal-orientated

activities and associated experiences of control, reward, social participation and support

(Siegrist & Theorell 2006). As the prevalence of unemployment is unequally distributed

across society, leaving those in lower socioeconomic positions at higher risk, this fact

contributes to the manifestation of a social gradient in health (see 3.1.1; Kasl & Jones 2000).

Second, wages and salaries provide the major component of the income of most people in

employment. There are substantial income inequalities in England (Jones et al 2008), leading

to material deprivation amongst the worst off. In addition, relative deprivation may be

experienced by people who are economically better off. Health-adverse effects of

inappropriately low income were demonstrated in several studies, thus adding further

evidence to the links between work, health and social inequality (see 3.1.4; Kawachi 2000).

Third, exposure to physical, ergonomic, and chemical hazards at the work place, physically

demanding or dangerous work, long or irregular work hours, shift work, health-adverse

posture, repetitive injury and extended sedentary work can all adversely affect the health ofworking people. Again, these conditions are more prevalent among employed people with

lower educational attainment and among those working in lower, less privileged occupational

positions (see 3.1.2; 3.1.3; Karasek & Theorell 1990).

Fourth, as the nature of employment and work has changed significantly over the last half

century, psychological and socio-emotional demands and threats evolving from an adverse

psychosocial work environment have become more wide spread in all advanced societies.

Technological progress and economic growth in the context of globalised markets and tradesresult in new types of tasks (e.g. information processing, personal services and service


6/86

- 6 -

centres). This has led to an unprecedented flexibility of employment arrangements and

contracts, often in combination with job instability and insecurity and with an increase in

work intensification and long hours of work. Todays economy inducing trade and financial

liberalisation carries a high risk of volatility and financial crisis, thus widening income

inequality, job instability and related material and psychosocial adversity (Blouin et al 2009).

This adversity includes conflicts within workplace hierarchies and power relations, restricted

participation of employees in decision-making, and a spectrum of covert or overt

discriminatory activities. Toxic combinations of these dimensions of work are frequent in

the current labour market, yet unequally distributed between occupations. Their highest

prevalence is found among the most deprived workers, specifically those in precarious jobs

defined by a lack of safety at work, by exposure to multiple stressors including strenuous

tasks with low control, low wage and high job instability (Benach et al 2000; Benach &

Muntaner 2007). As documented below (see 3.1.4) there is ample evidence on adverse effects

on health and well being produced by these conditions.

Overall, a social gradient of health-adverse employment and working conditions has been

documented in advanced societies, including England, leaving those in lower socio-economic

positions at higher risk (see 2.1; 3.2). Conversely, health-promoting and health protective

good working and employment conditions are more often experienced by people with higher

socioeconomic status who also enjoy better health and well being. The common scientific

approach towards studying associations of work with health and well being tends to focus on

pathogenic rather than salutogenic or protective and health-promoting effects, thus to

study adversity among lower socio-economic groups of employed people rather than

opportunities of good health and positive development. This is due to the primacy of the

consensual goal of reducing modifiable inequalities. In this report, too, emphasis is put mainly

on pathogenic aspects of work and employment. However, as will be documented, respective

scientific evidence provides a convincing basis of knowledge from which recommendationson how to develop and implement good (i.e. health-promoting and health-protective) work

can be derived. In particular, this will become evident in the context of theoretical models that

identify specific components within the complexities of work and employment that are of

critical importance for health (see 3.4.1).

While work and employment make a significant contribution to social inequalities in health, it

is equally true that obtaining employment and attained work-position are to a considerable

extent determined by exogenous factors over the life course (parental influences, educationalattainment, socio-economic position, peer groups and social networks, geographic location,


7/86

- 7 -

ethnicity and macro-economic environment). For this reason, the health impact of work and

employment cannot be assessed in isolation from the individuals wider social environment

(Kuh & Ben Shlomo 2004).

1.2. A conceptual approach towards disentangling links between employment relations

and health inequalities

In spite of growing scientific evidence regarding the effects of employment conditions on

health, few conceptual approaches have been proposed to disentangle the complex

relationships and pathways connecting employment conditions with health inequalities. One

such approach has been elaborated as part of the work of the WHO Commission on Social

Determinants of Health (Benach et al 2009) (Figure 1).

Figure 1. Conceptual approach towards analysing employment and working conditions withrespect to social inequalities in health

Healthinequalities

PoliticalPower

Government(Parties)

PoliticalPower Market(Unions, corporations,

Instituti ons)

PoliticalPowerSociety

(NGOs,

Community

Associations)

Power relations

Policies

Welfare state(Social {Policies)

Labor market(Labor Regulations,

Industri al Relations)

Social Class, Gender, AgeEthnicity, Migrant Status

Unemployment

Full employment

Standard/ Precariousemployment

Partially i nformalFully informal

Slavery / Child labour

Employment

Material deprivation &economic inequalities

Social & family

networks

Health Systems

Meaning of the arrows represented in the model:

Influence Mutual influence Interaction or buffering Influence at various levels

WorkOrganisation

Physio-pathologicalchanges

Health related

behaviours

(Life style /

medication)

Psychosocial

factors

WorkingConditions

Exposures and ris k

factors:

Injuries

Ph ysical & Che mical

Hazards

ErgonomicsPsychosocial

In this approach, macro-, meso- and micro-levels are linked, starting with governmental (e.g.

welfare regimes, social policies) and labour market conditions as upstream macro-

socioeconomic determinants. National labour markets are stratified according to power and

privilege (centre and periphery), socio-economic position, gender, race/ethnicity, migrantstatus and age. They include standard and non-standard work arrangements (e.g. informal,


8/86

- 8 -

unstable, temporary, seasonal work). Labour markets vary by stage of economic development,

sector and geographical region. At the meso-level, employment conditions influence health

both directly and indirectly, mediated by exposure to adverse working conditions. These

influences are partly non-specific or general, increasing the susceptibility of workers to a

range of different disorders, partly specific (e.g. occupational diseases). Downstream

pathways to the health of individual workers are mediated by noxious physical and chemical

hazards, psychosocial adversity expressed by psycho-biological mechanisms, and health-

related behaviours. These processes are not uni-directional but characterized by interaction

(e.g. effect modification) and feedback. Importantly, working people suffering from ill health

are faced with additional challenges of adaptation and integration into prevailing employment

and working conditions that may further affect their health in positive or negative ways.

Finally, the conceptual approach emphasizes the strong links that exist between work and

non-work (social and family life) conditions.

In this review, main emphasis is put on the ways of how employment and working conditions

affect health inequalities and how evidence of existing pathways can be used to help improve

the health of working populations. However, the current state of research on work and health

does not yet adequately address the complexities suggested in this conceptual approach

(Figure 1).

1.3. The process of knowledge generation

Any review of employment, working conditions and health faces various challenges. First,

much of the research into the interactions between these factors does not focus on health

inequalities and their causes. Second, knowledge on best practices and examples of policy

successes in lessening health inequalities is limited. Third, some of the most adverse working

conditions are often hidden or less well-known. Standard systematic reviews are limited by

the fact that studies are selected on the basis of the quality of the methods rather than ontheoretical considerations. The classical paradigm of randomised controlled trials is not

applicable in this context. Thus, following the precedent set by the Global Commission on the

Social Determinants of Health, we have taken a broader view of what constitutes evidence in

this field of scientific inquiry (Kelly et al 2006, Marmot & Friel 2008). Employing a wide

range of strategies of inquiry, a variety of methods, and multiple sources of data and evidence,

we synthesised the inputs of several disciplines.

First, in searching scientific literature we used digital bibliographic databases includedMedline, PsycInfo, Sociological Abstracts, Social Sciences Abstracts, EconLit, American


9/86

- 9 -

Business Inform, Business Abstracts, Public Administration Abstracts, Political Science, and

Worldwide Political Science Abstracts. Search strategies and key words were identified after

a series of tests and qualitative evaluations of each of the listings obtained. All searches were

limited by year of publication, from 1990 to 2009.

The information presented in Tables 1 and 2 (Appendix) is based on search strategies using

the following key words: Job insecurity, job instability, job loss, downsizing, temporary

employment, flexible work, non-permanent work, longitudinal study, cross-sectional study,

case control-study, psychological distress, psychosomatic symptoms, minor psychiatric

morbidity, self-rated health and health related behaviours. The information presented inTable

3 (Appendix) is based on search strategies using the following key words: demand-control

model, job strain model, decision latitude, job control, job demands, social support at work,

effort-reward imbalance model, over commitment, esteem, promotion prospects, job security,

coronary heart disease, cardiovascular disease, depression, physical and mental functioning,

self-rated health, stress-related disorders, cohort study, observational study, prospective study

and longitudinal study.

In evaluating study findings priority was given to results reported from prospective

observational studies in occupational health epidemiology as this study design represents the

gold standard in this field of research (see Table 3(Appendix)). This is due to the fact that

exposure assessment precedes disease onset, that the risk of disease is estimated as a function

of exposure and that effects can be adjusted for relevant confounding factors in a multivariate

analysis.

Second, we aimed at identifying relevant materials such as books, reports, and unpublished

documents. To identify and select on-line documents, we followed two main strategies: a)

Using metasearchers (ixquick.com, metacrawler.com, search.com) and a search engine

(google.com), searches were made for each employment dimension using key words. For

each dimension, additional key words were considered to limit results of the search to thetopic of interest. To reduce the number of documents we focused on publications after 1999-

2000.

Third, we consulted several key websites of relevant organisations, including non-

governmental organisations. Finally we had extensive personal communication with scientists

and other experts engaged in work and health related activities, so as to include their most

recent insights.

Despite our efforts in following these principles we cannot claim to represent all relevantaspects of this large and diversified field of research in this short report. Moreover, the


10/86

- 10 -

evidence corroborated through this process of knowledge generation may be biased to some

extent in terms of the language of the literature reviewed (largely English) and in terms of

publication bias (more positive than negative findings being published). A more elaborate

synthesis of information, e.g. by conducting a meta-analysis, was not feasible, given the

heterogeneity of sources, indicators, study designs and study populations available.

Nonetheless, with the primary interest being on evidence-based recommendations, in

considering those aspects of social inequalities in health that are attributable to work and

employment and, our aim was to contribute to the development of the evidence base for

good and sustainable work.

1.4. Defining a health-adverse psychosocial environment

As the nature of employment and work has changed substantially during recent decades,

psychosocial adversity at work has become a major concern of research and policy related to

work and health. This highly debated topic requires further conceptual clarification before it

can be integrated into evidence-based recommendations for policy.

An adverse psychosocial environment at work cannot be identified by direct physical or

chemical measurement. Theoretical concepts are needed to delineate particular stressful job

characteristics so that they can be identified at a level of generalization that allows for their

use in a wide range of different occupations. These concepts can be translated into measures

with the help of social science research methods (standardized questionnaires, observation

techniques, etc.) that meet the criteria of adequate reliability and validity of data collection. A

variety of concepts that encapsulate adverse psychosocial work environments have been

developed in occupational health psychology and sociology, social epidemiology and

organisational sciences (for reviews, see Antoniou & Cooper 2005, Cartwright & Cooper

2008). However, only a few have been tested with convincing study designs (e.g. longitudinal

observational investigations of initially healthy employed populations) and have addressed thesocial gradient in work and health. Among these, two models have received special attention,

the demand-control model and the effort-reward imbalance model.

The demand-control model (Karasek 1979; Karasek & Theorell 1990) posits that stressful

experience at work results from a distinct job task profile defined by two dimensions, the

psychological demands put on the working person and the degree of control available to the

person to perform the required tasks. This latter dimension is labelled decision latitude. Jobs

defined by high demands in combination with low control are stressful because they limit theindividuals autonomy and sense of control while generating continued pressure (high job


11/86

- 11 -

strain). Under these conditions, following the experience of control and mastery, it is

expected that excessive arousal of the autonomic nervous system would occur without any

compensatory relaxation response. Conversely, active jobs are expected to be health-

protective as they are defined by challenging demands that go along with a high degree of

decision latitude and learning opportunities, enabling individuals to experience positive

stimulation, success and self efficacy. A third dimension, social support at work, was added to

the original formulation. In this formulation, the highest level of strain would be expected in

jobs that are characterized by high demand, low control and low social support at work or

social isolation (iso-strain jobs) (Johnson & Hall 1988). Extensive tests of the demand-

control- (support) model showed that the concept, in its fully developed form, does not always

predict poor health but that this is more often the case if single components are analysed (see

3.1.4andTable 3 in Appendix).

A complementary model, effort-reward imbalance, is concerned with stressful features of the

work contract (Siegrist 1996). This model builds on the notion of social reciprocity, a

fundamental principle of all types of transactions that are characterized by some form of

utility. Social reciprocity lies at the core of the work contract which defines distinct

obligations or tasks to be performed in exchange with adequate rewards. These rewards

include money, esteem and career opportunities (promotion, job security). Contractual

reciprocity operates through norms of return expectancy, where effort spent by employees is

reciprocated by equitable rewards from employers. The effort-reward imbalance model claims

that lack of reciprocity occurs frequently under specific conditions. Failed reciprocity, in

terms of high cost and low gain, elicits strong negative emotions and associated stress

reactions with adverse long-term health consequences (see 3.1.4. and Table 3 in Appendix).

High cost-low gain conditions at work occur frequently if employed people have no

alternative choice in the labour market. This is often the case among those with low socio-

economic position or low level of skills, among elderly workers and, more general, in a highlycompetitive labour market.

These two models complement each other by focusing on toxic components of job task

profiles and employment contracts respectively. Low control and low reward are assumed to

be equally stressful experiences in the context of work that requires high levels of effort. They

both elicit negative emotions and enhanced stress responses with adverse long-term health

consequences. Thus, the co-occurrence of low control and low reward in demanding jobs has

been shown to increase the probability of ill health above and beyond the risk associated withexposure to the separate components (Peter et al 2002).


12/86

- 12 -

As mentioned earlier, the focus we have placed on these two models is due to the fact that an

accumulated body of empirical evidence on health-adverse effects is available from which

specific recommendations can be made (see 3.3.5 and 4). In addition, they contribute to the

role of work and employment in explaining social inequalities in health (see 3.2). Further

conceptual approaches were proposed and tested, indicating a growing awareness of the

importance of understanding health-damaging and health-promoting aspects of modern work

and employment. One such approach is concerned with organisational justice (Greenberg &

Cropanzo 2001), where inaccurate decision-making procedures and unfair treatment by

supervisors contribute to poor health (Elovainio et al 2002). Despite some potential overlap

with the above mentioned approach, the models of effort-reward imbalance and organisational

justice have been shown to predict health outcomes independently (Kivimki et al 2007).

An additional multidimensional construct, employment precariousness, has been proposed to

capture more distal, labour market and related determinants of health (Benach & Muntaner

2007).

1.5. Structure of the report

In this report it is argued that there is substantial evidence that some of the employment and

working conditions in England are associated with adverse health outcomes (see section 2.1).

Many of these conditions are socially patterned, leaving those in lower positions at higher

risk. The extent to which inequalities in adverse work and employment contribute to health

inequalities is still debated in current research. At least two concurrent explanations exist.

One such explanation claims that social selection accounts for a large part of work-related

social inequalities in health. This means that people with unfavourable socioeconomic,

psychosocial or biological background are more likely to end up in stressful jobs in adult life

(Macleod et al 2001, Nettle 2003). In the context of a life-course approach to chronic disease

development this argument is important. However, several epidemiological studies havecontrolled for adverse childhood circumstances and found that work-related factors have a

stronger explanatory power than social background-related factors (Brunner et al 2004,

Marmot et al 2001, Melchior et al 2006, Kivimki et al 2005). For this reason, it seems

appropriate to explore the extent to which stressful work and employment follows a social

gradient and contributes towards explaining social inequalities in health. As a consequence,

targeting these work and employment conditions through prevention or intervention provides

a plausible approach towards reducing these inequalities.


13/86

- 13 -

The recommendations of this report (see section 4) are derived from scientific knowledge that

evolved mainly from tackling the following four key questions:

First: Is there sufficient evidence of a causal link between adverse work and employment

conditions and reduced health? This question is addressed insection 3.1

Second: Is there sufficient evidence that adverse work and employment conditions mediate

the association between socioeconomic position and health? This question is addressed in

section 3.2.

Third: Is there sufficient evidence that work- and employment related interventions improve

health and, thus, may contribute towards reducing social inequalities in health? This question

is addressed in section 3.3.

Fourth: Is there sufficient evidence that favourable economic effects result from work-and

employment-related interventions? This question is addressed in section 3.4 where the role of

broader socio-political and economic conditions is also discussed.

In section 4, recommendations for action, both general and specific, are derived from answers

to these questions. Links between these recommendations and reported evidence are indicated

by cross-references.

Given the central importance of work and employment in adult life and its close links with

personal, family and civic life, it is not feasible to address all aspects relevant to health in

detail in this report. In particular, the evidence on several topics has not been summarised -

including the working conditions of self-employed people and their effects on health (Saarni

et al 2008), the cumulative or protective effects of a work-life balance (or imbalance) and

their effects on unequal health (Artazcoz et al 2004, Westman 2002). Gender-specific

variations in the associations between work, social inequality and health are also important,

but are not considered in detail in this report (Messing & Silverstein 2009, Weidner et al

2002).

As this report draws on the international state of the art in this field of research, a criticalquestion concerns the relevance of available knowledge to the English context. While this

question is dealt with throughout the report, the next section (section 2) is specifically focused

on a discussion of employment, working conditions and related policies in England.

2. The context for recommending future policies and interventions

2.1. Employment and working conditions in England

As is the case for many countries, the labour market in England has been significantlyaffected by globalisation and related financial and economic issues. Economic, financial and


14/86

- 14 -

trade decisions of large corporations operating at a global scale affect working conditions of

large parts of the workforce, labour standards, occupational health and safety regulations,

wages and measures of social protection. Currently, the most visible developments concern

rising unemployment, an increase in non-standard employment, work intensification and the

growth in numbers in vulnerable groups.

According to the most recent National Statistics on the labour market (Office for National

Statistics 2009) the employment rate at working ages was 74.1 per cent in the three months to

January 2009, with an unemployment rate of 6.5 per cent. As a consequence of the current

financial crisis, 266,000 people were laid off in that three-month period, raising

unemployment above 2 million for the first time since 1997 and affecting all economic

sectors. The increase in unemployment rates, compared to a year earlier, exceeded two

percentage points in the North East, Yorkshire and the Humber, the West Midlands and

Wales. An earlier analysis showed that unemployment rates in the previous three months

were highest for those who had worked in elementary occupations (9.7 per cent) and sales and

costumer service occupations (7.5 per cent). Rates were lowest for professional and

managerial occupations (1.6 and 2.2 per cent). A report by Oxford Economics estimates that

more than a million British workers will lose their jobs over the next two years (Taylor 2009).

2.1.1 Recent changes in the workforce composition

Large-scale enterprises account for the main share of total employment (46 per cent) in the

UK economy. While the number of large, multinational companies has traditionally been

high, economic restructuring during the 1980s reduced the average size of workplaces and

encouraged the growth of small enterprises. Moreover, the proportion of own account self-

employed workers has increased (Walters 2004).

The changes in the structure and organisation of work and labour markets are important

influences on the practicability of workplace arrangements for representing workers interests

in improving occupational health. Outsourcing and downsizing by large firms has contributed

to growing job insecurity, self-employment, and precarious employment. These practices

have been driven by management strategies such as lean production, flexible work and

engineered standards. The growing influence of neo-liberal policies in government has led to

practices such as privatisation and competitive tendering, with these developments being

experienced in both private and public sectors (Walters 2006).


15/86

- 15 -

The relationship between the labour market and health has also been influenced by changes

introduced through legal frameworks addressing employment security, industrial relations and

social welfare. Workers representation in these activities seems less pronounced in the UK

compared to many Western European countries where intermediary institutions between state

and single enterprises are more strongly developed. For instance, a recent survey indicates

that one third of employees only have access to some type of collective bargaining (Parent-

Thirion et al 2007).

The last 25-30 years has witnessed some fairly radical changes in the British labour market.

The continuing decline of the manufacturing sector has given rise to the dominance of the

service sector leading to changes in both the industrial structure of employment and

occupational composition. Until employment rates started to decline in the middle of 2008,

they had increased fairly steadily over the previous 15 years. However, in 2008 the

employment rate was roughly the same as it had been in 1971. A more detailed examination

of the data shows a very different picture for men and women. Between 1971 and 2008 male

and female (working age) employment rates show considerable convergence; increasing by 14

percentage points for women and falling by 13 percentage points for men. This meant that in

April 2008, 79 per cent of working age men and 70 per cent of working age women were in

employment (McKnight 2009). Women are much more likely than men to work part time (40

per cent compared with 10 per cent), mainly to fit employment around caring for their

children. These jobs are generally of lower quality (in terms of pay, conditions and status)

than full time jobs and this relates to the overall lower position of women in the labour

market. Significant progress has been made in narrowing the pay gap between male and

female employees but it remains high (particularly for women working part time).

The 1980s and 1990s recessions and the restructuring associated with the decline in

manufacturing all took their toll on male employment. This was particularly the case for older

male workers for whom employment rates among men aged 50-64 fell from around 90 percent in the early 1970s to less than 65 per cent in the early 1990s (Pensions Commission,

2005). Not only did unemployment increase but inactivity rates among working age men

increased too. This was coupled with increases in rates of disability among the working age

population and lower rates of economic activity among this group. Union membership

continued its steady long term downward trend through the 1990s and 2000s (from 32.5 per

cent in 1995 to 28.4 per cent in 2006) (Grainger & Crowther 2007).

Additional relevant changes concern the age composition of the workforce, the increase ofmigrant workers and changes in educational level. The employment rate of people aged 50


16/86

- 16 -

and over has increased steadily since 1992. For those aged 50 to state pension age, the rate

increased by 8.6 percentage points and for those individuals of state pension age and above,

the increase was 3.6 percentage points. This resulted in an increasing number of older people

in the workforce.

An increase in migrant workers has been observed in recent years, in part due to the

enlargement of the European Union in 2004 (Ker &Kahn 2009). In the first quarter of 2009,

about 13 per cent or 3.8 million employed people were born overseas (compared to 2.0

million in 1997 and 2.6 million immediately before EU enlargement in 2004). Of those born

abroad, numbers born in the EU increased from 694 thousand at the time of enlargement to

1.2 million in 2009. The new EU states contributed 76 thousand and 518 thousand,

respectively, in 2004 and 2009 (ONS 2009). A majority of those from new EU states (55 per

cent) were employed either in elementary occupations or as process, plant and machine

operatives, compared to 18 per cent of the UK born (Ker &Kahn 2009).

Partly in response to the shift from a manufacturing to a service led economy (where

academic qualifications have greater value over vocational skills), the UK working age

population has become more academically qualified. The proportion without any educational

qualifications has been reduced by a third over the last ten years.

However, increases in higher education over the 1990s were disproportionately enjoyed by

the most advantaged. Higher education participation rates increased from 35 per cent to 50 per

cent for young people from non-manual backgrounds but from 11 to 19 per cent for young

people from manual backgrounds. Some recent evidence suggests that this socio-economic

gap narrowed between 2002 and 2006 (DIUS 2008). The higher qualified enjoy a much more

advantaged position in the labour market with higher earnings and lower incidence of

unemployment (Palmer et al 2008).

To monitor social inequalities in current work and employment conditions in a reliable and

valid way, a new National Statistics Socio-economic Classification (NS-SEC) was introducedin 2001. This classification was the first to have relations and conditions of employment as its

conceptual basis where it is assumed that social power arises from the division of labour. The

allocation of detailed occupation groups to classes was based on theory and empirical

investigation. A series of specially commissioned questions, designed to capture the

theoretical dimensions of the conceptual basis, were included in the UK Labour Force Survey

to assist with the allocation of occupation groups. The five key dimensions were the structure

of pay, period of notice required, promotion prospects and flexibility in working time(autonomy). In terms of its theoretical basis, NS-SEC represents a structural model where


17/86

- 17 -

individuals occupy social class positions that shape their lives and determine a variety of

outcomes, one of which is health (Rose & Pevalin 2003).

One of the biggest challenges in terms of tackling the social gradient in health is that some of

the social gradients in the social determinants of health inequality have considerably widened

over the recent past. Social change over the last few decades in the UK has meant that there

has been considerable growth in managerial and professional occupations (more room at the

top) and a relative reduction in a range of occupations at the intermediate and lower end of

the social scale. This has resulted in an increase in the share of employees working in jobs

with better relations and conditions of employment with health outcomes among employees

expected to improve. However, residualisation has meant that those at the lower end of the

social scale are increasingly more disadvantaged across a range of dimensions with the

social gradient in health outcomes expected to have increased as a result.

Even if there were no change in health outcomes, a simple redistribution of the population

across the social scale, with the most advantaged become more heavily concentrated in the

top social class and those at the lower end becoming a more homogenous group of very

disadvantaged employees, would result in an increase in health inequality. Layered on top of

this redistribution, some determinants such as earnings inequality and the experience of

unemployment have become considerably more unequal between employees along the social

gradient (Annan 2009, Dickens & McKnight 2008a). Inequality in hourly, weekly, monthly

and annual earnings became considerably greater over the 1980s and to a lesser extent over

the 1990s. In terms of annual earnings, recent research evidence has shown that the highest

earning 10 per cent of employees in 1980 were earning approximately 10 times the lowest

earning 10 per cent. This increased to around 17 times by 1990 and 20 times by 2000.

Inequality in annual earnings, according to this measure, doubled over this 20 year period

(McKnight 2009). While earnings form the largest component of household income for the

majority of the working age population, the impact of benefits and taxation is also important.However, a recent study found little net redistributive effects of these over the last 30 years

(Jones et al 2009).

Work to validate NS-SEC showed a clear relationship between the experience of

unemployment and social class (both short term and long term) even after controlling for

gender, age and household circumstance. Individuals occupying labour contract class

positions were found to be considerably more likely to experience unemployment than those

occupying service relationship classes (Elias & McKnight 2003).


18/86

- 18 -

In such an environment it is clear that, in assessing the likely impact of policies designed to

reduce health inequalities, it is necessary to take into account changes in the social context

and have a good understanding of the underlying changes in the labour market.

2.1.2 Critical employment conditions

The majority of the English workforce is employed on full-time contracts. While the country

is known for its long working hours culture average weekly hours of employment fell from

33.5 to 32 between 1995 and 2005, a similar decrease in percentage point was also recorded

among full time employees. The share of employees working very long hours, 48 hours or

more a week, also fell by 20 per cent. This trend was no doubt related to the Working Time

Directive which partly came into force in 1998, although the limit on weekly hours of work

remains voluntary. Despite the high prevalence of full-time work, part-time work, temporary

work and other types of non-standard work are relevant to the topic of this report.

Temporary working, although heterogeneous, is generally considered to be of lower quality

than permanent employment. Concern grew over the 1990s as the share of employees

working in temporary jobs increased. Yet, in 2008, the rate fall to 5.5 per cent. Various

policies are being designed to improve the conditions of employment of temporary workers

(e.g. the Agency Workers Directive).

The percentage of individuals working part time increased from 23.6 per cent in 1992 to 25.5

per cent in 2008. While this results from free choice in many employees (women, older

workers), male part-time employees often report that they could not find a full-time job (Kent

2009). A recent survey found that the UK had one of the lowest proportions of people holding

indefinite contracts, and that there were a large number of workers without any contract

(Parent-Thirion et al 2007). However, all UK employees are covered by a range of Statutory

Rights. These include the statutory right to a written contract of employment and rights that

cover unfair dismissal, paid holiday entitlement, redundancy pay and, for fixed termemployees, having the same contractual rights as comparable permanent employees.

Many forms of non-standard work arrangements and precarious jobs such as contingent,

unregulated underground or home-based, are characterised by variable schedules, reduced job

security, lower wages, hazards at the workplace and stressful psychosocial working

conditions. Workers having a permanent contract have more skills and credentials, have more

information of the workplace hazards, experience less hazardous work conditions, and have

better health outcomes. Research has also shown that when workers have less skills andcredentials, they also tend to experience hazardous working conditions, including physical


19/86

- 19 -

strain, low job control, greater noise and air pollution, shift work, a monotonous job, and a

hectic work pace, as well as worse self-reported health and a large number of health outcomes

(Vahtera et al 1999; Schrijvers et al 1998; Siegrist & Marmot 2006).

Informal employment is an important type of precarious employment varying according to

type of production unit and type of job. Type of production unit is defined in terms of the

legal organization and other enterprise-related characteristics, while type of job is defined in

terms of employment status and other job-related characteristics. Production units are

classified into three groups: formal sector enterprises, informal sector enterprises, and

households. Jobs are distinguished according to status-in-employment categories and

according to their formal or informal nature. For employment status, different types of groups

can be described: own-account workers; employers; contributing family workers; employees;

and members of producers cooperatives.

Workers holding informal jobs are disadvantaged compared to formally hired workers in

several aspects that separately or together affect their occupational health. The most important

factor is poverty, since several studies show that firms in the informal economy usually have

low profits, and informal workers have lower salaries than those in formal firms. In addition,

individuals working in the informal sector are unlikely to be making National Insurance

contributions. For this reason, it is less likely that they will be building up entitlement to

protective, contribution-based benefits which they could draw on when they are unemployed,

sick, pregnant and when they retire. This makes their position precarious both now and in the

future. Informal employment is relatively prevalent among young people and, overall, is

higher in the UK than in many other European countries (Stanculescu 2005). The scientific

literature on occupational health and the informal economy is scarce and most studies are

descriptive, a fact that limits the generalisation of results (da Silva et al 2006; Fongichigong

2005; Hernandez 1996; Nilvarangkul et al 2006; Lowenson 1998; Rongo et al 2004; Iriart et

al 2006; Santana & Loomis 2004; Gutberlet & Baeder 2008).Child labouris a type of work that is of special concern. Although quantitative estimates are

not precise, studies conducted over the last two decades in the UK conclude that the extent of

child labour is still an issue (Pettitt 1998; ODonnell & White 1998; Somerset 2001).

Although the Government introduced the Childrens Bill in 2004, designed in part to extend

the protection offered to children, child work continues to require careful monitoring as child-

labour is often associated with problems related to the physical, physiological, mental and

social development of children as well as accidents (Health and Safety executive 2003). Childlabour may also directly compromise height, which can be regarded as a biological indicator


20/86

- 20 -

of social injustice, and recently seen as a relevant component of the so-called physiological

capital (Eijekemans et al 2005; Gunnarsson; Dantas 2005; Duyar & Ozener 2005; Fassa 2003;

Yamanka & Ashworth 2002; Fogel 2006; Hawamdeh & Spencer 2003). Several recent policy

documents have addressed issues of special concern about the development of child labour

(UNICEF 2003a, 2003b, 2004; TUC/NSPCC 2001).

Migrant workers are another group at potential risk. A report, undertaken in South

Lincolnshire in 2005 on behalf of the East Midlands Development Agency (Zaronaite &

Tirzite 2006) identified a range of problems faced by migrant workers due to exploitative

practices of employment agencies and developed a set of key recommendations in relation to

exploitation at work, exploitation in accommodation, and education, training and integration

(Palmer et al 2008). Remittance payments put an additional strain on many migrants, leading

them to work long hours, hold multiple jobs, live in overcrowded accommodation and eat a

poor diet (Datta et al 2006). However, it is difficult to generalise as there is a considerable

amount of variation in the chances of migrants being employed and their relative earnings by

country of birth (Dickens & McKnight 2008b; Rutter & Latorre 2009).

A minority issue concerns slavery. A report on the UK position highlighted the circumstances

of those working in highly exploitative conditions, with no rights and threatened with the fear

or reality of violence (Craig et al 2007). Studies reveal abuse and exploitation of migrant

domestic workers (Oxfam & Kalayaan 2008).While there are no reliable figures on the scale

of forced labour in the country, there is a range of qualitative accounts on the subject, based

both on first-hand and anecdotal accounts of foreign nationals trafficked into the UK.

Women from poorer countries who do not find formal or informal jobs in economic sectors

such as agriculture/horticulture, contract cleaning and residential care (Anderson & Rogaly

2005) are particularly at risk of trafficking for sexual exploitation (Skrivankova 2006). No

valid figures are available on numbers involved and whether (and to what extent) this

trafficking has involved children. The working and living conditions of these vulnerablegroups and the limited statutory protection afforded to them warrants particular attention

(Anderson & OConnell 2003, Matthews 2006). For example, the UK has yet to sign and

ratify the Council of Europes Convention on Action Against Trafficking in Human Beings,

or to ratify the UN Palermo Protocol (Craig et al 2007).


21/86

- 21 -

2.2. Recent initiatives on improving health and work in EnglandIn this section we briefly describe some of the recent policy initiatives designed to tackle

unemployment, inactivity and other aspects of poor relations and conditions of employment

which are known to be associated with poorer health outcomes. A number oflabour market

policies introduced over the last ten years have potentially improved health among the work

force. In terms of reducing unemployment, the New Deal programmes were designed to

tackle high unemployment rates among specific groups (e.g. long term unemployed, young

people, over 50s, lone parents). In the first few years after this programme was introduced

evaluation evidence suggests they have had a modest but significant impact on helping

unemployed people into work. Different groups faced differing degrees of compulsion to

participate in the programmes which involved active assistance from personal advisors to

prepare for and seek employment. Young people (18-25 years) who had not found work after

a set period faced benefit sanctions if they did not take up subsidised employment or a

restricted number of work related options (Blundell et al 2003).

For the first time lone parents were offered voluntary assistance to prepare for and find work.

The New Deal programmes have evolved over time and elements have now been integrated

into standard conditions for all job seekers. In addition to activating the unemployed into find

work, there has been a second strand of policies designed to make work pay. In-work

benefits have been extended to include a much wider group of low paid employees and these

Tax Credits are more generous than those previously available. Tax credits increase the

financial incentive for individuals with low earnings potential to find and remain in work. In

1999 for the first time a national minimum wage was introduced and various other changes to

the tax and national insurance schedules led to increases in the net earnings of low paid

workers.

More recent developments have shifted the focus onto those who have traditionally not

received work search assistance and whose out of work benefit entitlement has not beenconditional on taking active steps to find work. Disabled people and lone parents have

traditionally experienced high rates of non-employment and high associated rates of poverty.

A number of initiatives have been tried to reach disabled people but the most significant was

introduced in October 2008 when the whole structure and conditions attached to claiming

benefit on the basis of disability were changed. All applicants now have to undergo a Work

Capability Assessment. If they meet various criteria individuals can qualify for the new

Employment and Support Allowance (ESA), if they do not then they have to apply for JobSeekers Allowance.


22/86

- 22 -

A number of policies have been designed to make it easier for individuals (particularly

women) to combine caring responsibilities with work. The National Childcare Strategy,

announced in 1998, aimed to make available accessible, affordable, good quality childcare for

all children aged 0-14 years (DfEE 1998). Some gains have been made in working towards

this objective with free pre-school places available for 3 and 4 year olds and an expansion in

child care places (Stewart 2009). However, deficiencies remain and coverage is not universal

with particular problems with certain types of childcare (such as wrap around care). Since

2003 employees have had a statutory right to request a flexible working arrangement as long

as they have been continuously employed with the same employer for a defined minimum

time. This applies to parents and legal guardians and those who care for a spouse or relative.

However, this does not confer the right to flexible working - only a statutory right to ask for it

and for their application to be given serious consideration.

Antidiscrimination legislation on the basis of sex introduced in 1975 was extended to include

age (2006), religion or belief and sexual orientation (2003). The Equality Bill before

Parliament in 2009, if passed, would consolidate and extend previous legislative on age,

disability, gender reassignment, marriage and civil partnership, pregnancy and maternity,

race, religion or belief, sex and sexual orientation. It would also Impose a duty on some

public authorities to have due regard to socioeconomic considerations in deciding their

strategic priorities (UK Parliament 2009).

Importantly for health inequalities employees on fixed term contracts now have the same

minimum rights as permanent employees in terms of pay and conditions, benefits packages,

occupational pensions schemes (but usually have to have a contract for 2+ years) and

protection against redundancy or dismissal. However, temporary workers (particularly agency

workers) conditions remain inferior to those enjoyed by most permanent employees and

further steps are being explored to improve their conditions of employment. The 1995

Disability Discrimination Act built on and extended earlier disability discriminationlegislation and required employers to make reasonable adjustments to ensure that they do not

discriminate against disabled customers or employees. It prohibits discrimination in relation

to employment of disabled people, including recruitment, training, promotion, benefits,

dismissal, etc.

In conclusion, recent initiatives on improving aspects of employment known to be related to

health outcomes such as, work participation and employment prospects, protection at work

and improving income of low paid workers - in England have produced some favourable


23/86

- 23 -

results. Yet, in view of the challenges described additional efforts are needed, in particular

those aiming at the reduction of social inequalities in health of the workforce.

3. Answering the key questions3.1. Health effects of adverse work and employment conditions: a selective review

3.1.1. Unemployment, job instability and temporary employmentThe impact of unemployment on health is well established in England. Unemployment

increases rates of depression, particularly in the young. Parasuicide rates in young men who

are unemployed are substantially higher than for those in employment (Dorling 2009).

Several epidemiological studies from England and Wales (Morris et al 1994, Moser et al

1987, Bethune 1997, Mitchell et al 2000), from Scandinavian and other countries

(Martikainen & Valkonen 1996, Iversen et al 1987, Pensola & Martikainen 2004, Voss et al

2004, Ahs & Westerling 2006, Kivimki et al 2003, Gallo et al 2004) document elevated risks

of fatal or non-fatal cardiovascular or cerebrovascular events or of all-cause mortality among

unemployedas compared to permanently employed people. Risks are particularly high among

the long-term unemployed.

The prospective study design, and the adjustment for baseline health observed in some of

these studies reduce the probability of reverse causation. Effect sizes are usually in the range

of 1.5 to 2.5. In the ONS Longitudinal Study, excess mortality from suicide was obvious

among unemployed men (Moser et al 1984). In the British Household Panel Study

unemployment among people in the most disadvantaged social group was related to elevated

risk of incident limiting illness (Bartley et al 2004). Other studies point to impaired mental

health, specifically depression, as a consequence of unemployment (Kasl & Jones 2000,

Kaplan et al 1987), whereas becoming depressive in turn increases the probability of future

unemployment and loss of income (Whooley et al 2002). Becoming re-employed is generally

associated with a reduction in symptomatology (Kessler et al 1989, Kasl & Jones 2000), but

not in the rate of mortality (Bethune 1997).

A substantial amount of evidence is now available on adverse health effects due to job

insecurity and its most important determinants, downsizing, restructuring and outsourcing. In

Table 1 (Appendix) the current state of art is summarised in a systematic way. It is of

particular interest to see that exposure to job instability due to downsizing was shown to

increase health adverse behaviour (Kivimki et al 2007), musculoskeletal problems (Kivimki

et al 2001), work-related health symptoms (Dragano et al 2005), sickness absence (Vahtera et


24/86

- 24 -

al 2004, Westerlund et al 2004a), disability pension (Vahtera et al 2005) and mortality

(Vahtera et al 2004) (for review Ferrie et al 2008). In this latter study from Finland all-cause

mortality was increased by about 40 per cent among those experiencing (and surviving) heavy

downsizing, and coronary heart disease mortality was increased by almost a hundred percent

(Vahtera et al 2004). Conversely, turning from insecure to secure employment was associated

with improved health (Virtanen et al 2003). In the British Whitehall II study, exposure to

threat of a major organizational change was associated with adverse changes in longstanding

illness, sleep patterns and minor psychiatric morbidity (Ferrie et al 1998). Other reports

document effects of job instability, in combination with subjectively perceived job insecurity,

on atherogenic risk (Siegrist et al 1988, Mattiassin et al 1990). Several studies indicate that

job insecurity, defined as the discrepancy between the level of security experienced and the

level they prefer, is related to poor mental health (Ferrie et al 2008). Job instability and

insecurity can also occur in rapidly expanding companies in addition to the more obvious

processes of downsizing. One Swedish study found elevated odds ratios of long-term sickness

absence and hospitalization of employees in the years following a period of rapid

organisational expansion (Westerlund et al 2004b). However, the current evidence on health

adverse effects of perceived job insecurity is mixed (see Table 1, Appendix and Muntaner et al

1998).

Temporary employmentor working on a fixed term contract, a condition that is common in

about 15 per cent of the workforce in Europe, is associated with increased risks of a variety of

adverse health outcomes. Again, the current state of the art is summarised in a systematic way

in a separate Appendix (in Table 2, Appendix).

Along with labour market flexibility, deregulation of labour markets around the world is a

potential source of deterioration in workers health in some of the new types of temporary

employment. Although a majority of studies suggest adverse effects on health, temporary

work is sometimes related to an improvement in health (Virtanen et al 2003), possiblyreflecting effects of different labour market regulations in different countries (Rodriguez

2002) or the heterogeneity of circumstances in which people take on temporary work.

The risks associated with temporary work included increased occurrence of alcohol-related

causes of death in both genders and an increase in smoking-related causes of death in men,

with hazard ratios ranging from 1.2 to 1.6 (Kivimki et al 2003). Mortality risks are

substantially stronger if temporary work is continued on an involuntary basis or in

combination with feelings of dissatisfaction (hazard ratios ranging from 2.1 to 2.6) (Ntti et al2009). In a review of 27 studies on health effects of temporary work most consistent


25/86

- 25 -

associations were found with regard to reduced mental health (Virtanen et al 2005a).

Additionally, increased risks of accident and musculoskeletal disorders were reported

(Benavides & Benach 1999, Silverstein et al 1998).

Although temporary work arrangements have tended to create jobs during periods of high

unemployment in most European countries, this can have a detrimental effect on health. Most

recent research has indicated that under equal working conditions, such types of employment

tend to be associated with several health problems (Benavides & Benach 1999; Benavides et

al 1999; Benavides et al 2000) such as distress (Cannuscio et al 2004), fatigue,

musculoskeletal disorders (Benavides et al 2000; Benach et al 2004), self-perceived health

(Virtanen et al 2005b; Artazcoz et al 2005), liver disease, mental disorders (Kim et al 2008),

absenteeism (Benavides et al 2000; Benach et al 2004, Virtanen et al 2006) and stress

(Benavides et al 2000; Virtanen et al 2006). Against this background, Dutch and Danish

governments have developed flexicurity labour markets, intended to expedite flexibility to

the benefit of employers, and at the same time, to give greater job security to employees.

3.1.2. Physical, ergonomic and chemical hazards at work

The European-wide panel survey on working conditions indicates that every sixth worker in

Europe is exposed to toxic substances at the workplace, and almost one third is exposed to

noise at work, at least intermittently (Parent-Thirion et al 2007). A detailed account of

occupational diseases cannot be given here (McDonald 2000), but it is evident that physical

and chemical stressors at work make a significant contribution to the burden of work-related

diseases and injuries (Verma et al 2002). Specifically, occupational groups with a high

percentage of workers in lower socio-economic positions are at elevated risk of occupational

injuries and accidents, such as construction workers, agricultural workers, transport workers,

or miners (Arndt et al 2005). Moreover, unhealthy or restricted posture at work, repetitive

movements and heavy lifting are more prevalent among lower status workers, and theseconditions increase the risk of musculoskeletal disorders (Bernard 1997). Workers exposed to

these physically stressful conditions are less likely to be able to work until retirement age

(Parent-Thirion et al 2007), and their risk of disability pension is increased by 50 to 100

percent compared to unexposed workers (Blekesaune & Solem 2005, Krokstad et al 2002,

Lund & Csonka 2003, Karpansalo et al 2002).

Physical, ergonomic and chemical hazards at work are often combined with an adverse

psychosocial work environment, thus multiplying health risks among exposed people. Few


26/86

- 26 -

studies have documented the long-term health effects of such cumulative exposures in any

detail (see Devereux et al 2002, Dragano 2007).

3.1.3. Shift work and other work time factors

The results of several epidemiological studies, suggest that the risk of cardiovascular disease

in shift workers is increased by about 40 percent compared to daytime workers (Hrm 2006,

Tchsen et al 2006, Haupt et al 2008, Ellingsen & Bener 2007). Similarly, an increased risk of

developing a metabolic syndrome was observed among shift workers, with a relative risk of

about 1.7 (de Baquer et al 2009, Karlsson et al 2001). Additional investigations demonstrate

an elevated risk of accidents, particularly among evening and night shift workers (Bambra et

al 2008a). Reported health effects are contingent on duration of shift work, with marked

increases after more than 10 years of continued exposure (Steenland 2000). However, there

are some inconsistent results, and the processes mediating the reported associations (sleep

disturbances and mismatch between circadian rhythms, disturbed work-life balance, changes

in health lifestyle) are still debated (Hrm 2006). Night shifts are particularly relevant as a

potential source of work accidents, cardiovascular and gastro-intestinal problems and

eventually cancer (Swerdlow 2003). Potential links between shift work, chronodisruption and

the pathogenesis of cancers are currently debated in international occupational health research

(Erren et al 2009).

In recent years, with increasing flexibility of work time patterns, studies documented adverse

effects on health produced by extended or irregular work hours. For instance, working more

than 11 hours a day is associated with a threefold risk of myocardial infarction (Sokeyima &

Kagamimori 1998, van der Hulst 2003), and a fourfold increased risk of type 2 diabetes

(Kawakami et al 1999). Moreover, in jobs with an overtime schedule the risk of injury is

increased by 61 percent among American workers (Dembe et al 2006). In an 11-year

longitudinal study among Finnish workers atherosclerotic plaque growth in the carotis wasproportional to number of days worked per week and to annual work hours (Krause et al

2009).

A further temporal factor concerns work time control. Low work time control is associated

with reduced health (Ala-Mursala et al 2004), whereas increased work time control moderates

adverse health effects of stressful work (Ala-Mursala et al 2005).


27/86

- 27 -

3.1.4. An adverse psychosocial work environment

High demand in combination with low control (job strain) and effort-reward imbalance at

work are associated with elevated risks of several highly prevalent chronic diseases in midlife,

as evident from longitudinal observational studies carried out over the past twenty years (see

Table 3, Appendix). To a lesser extent, this holds true for a third concept, organisational

justice. The strongest available evidence of associations of the two work stress models, job

strain and effort-reward imbalance with adverse health is summarised in Table 3

(Appendix). Concerning cardiovascular disease, a majority of at least 20 reports derived from

prospective studies document elevated odds ratios of fatal or non-fatal cardiovascular (mostly

coronary) events among those reporting job strain or effort-reward imbalance (Belkic et al

2004, Eller et al 2009, Kivimki et al 2006, Marmot et al 2006). Overall, risks are twice as

high among those with job strain or effort-reward imbalance compared to those who are free

from stress at work. Effects are stronger in men than in women and more pronounced in

middle-aged than older working populations. Similar effects are observed in case of re-

infarction after survived first coronary heart disease (Aboa-boul et al 2007). Two reports

based on the concept of organisational justice demonstrate an elevated cardiovascular risk

(Kivimki et al 2005, Elovainio et al 2006).

In addition several cardiovascular risk factors are associated with an adverse psychosocial

work environment in terms of job strain and effort-reward imbalance, in particular metabolic

syndrome (Chandola et al 2006), type 2 diabetes (Kumari 2004), hypertension (Schnall et al

2000), elevated fibrinogen (Brunner et al 2004, Vrijkotte et al 1999), atherogenic lipids

(Siegrist et al 1988), obesity (Kivimki et al 2002), health-adverse behaviours (Head et al

2004, Siegrist & Rdel 2006) and markers of dysregulated autonomic nervous and endocrine

system activity (Chandola et al 2008, Hintsanen et al 2005, Vrijkotte et al 2000, Steptoe et al

2004).

A second, widely prevalent chronic disorder, depression, is associated with stressful work.The large majority of results from 12 prospective investigations confirm elevated risks of

depression among employees with job strain or effort-reward imbalance or a co-manifestation

of both, and odds ratios vary between 1.5 and 3.6, depending on type of measure, gender and

occupational group under study (Bonde 2008, Siegrist 2008). Again, psychobiological

pathways that may trigger affective disorder were analysed with regard to job strain and

effort-reward imbalance, especially so dysregulated patterns of cortisol secretion (Chandola et

al 2008, Bellingrath & Kudielka 2008) and endogenous inflammation (Hamer et al 2006).


28/86

- 28 -

Other health outcomes significantly related to job strain or effort-reward imbalance concern

reduced physical and mental functioning (Stansfeld et al 1998), musculo-skeletal disorders

(Bongers et al 2002, Gillen et al 2007, Rugulies & Krause 2008), sickness absence (Marmot

et al 2006, Head et al 2007) and disability pension (Blekesaune & Solem 2005, Dragano

2007, Stattin & Jrvholm 2005).

It should be pointed out that many of these new findings were obtained from the Whitehall II

study in England where the concepts of demand, control and support at work, effort-reward

imbalance at work and organisational justice have been tested extensively, with a variety of

health outcomes and intermediary markers (Table 3, Appendix).

In conclusion, despite the fact that several studies have reported negative results, there is a

substantial body of scientific evidence on the health effects of an adverse psychosocial work

environment. This provides a solid basis for developing a range of work and employment-

related interventions (see Section 3.3).

3.2. The role of work and employment in explaining social inequalities in health

As indicated above, the distribution of adverse job conditions across working populations

(unemployment, job instability and temporary employment, physical, ergonomic and

chemical hazards, shift work and other work time factors) is strongly related to the social

gradient. Those in more disadvantaged groups are more often exposed than those in more

privileged positions. Does the same hold true for an adverse psychosocial work environment?

While this question has not yet been thoroughly researched, the components low control at

work (job strain model) and low reward (effort-reward imbalance model) were repeatedly

found to follow a social gradient in the expected direction (Bosma et al 1998, Brunner et al

2004, Marmot et al 2006, Niedhammer et al 2000). However, the prevalence of demand,

effort and over commitment is often higher in higher occupational status groups (Karasek

et al 1998, Siegrist et al 2004), resulting in a mixed pattern of evidence. Nonetheless, in arecent comparative study of adverse working conditions among 50 to 65 year old employees

in 11 European countries, a consistent social gradient of effort-reward imbalance and low

control at work was observed (Siegrist et al 2009).

Using multivariate regression analysis to test the mediation hypothesis, some evidence to

support this hypothesis has been found. In the Whitehall II study, low control at work was

independently associated with incident coronary disease and with low socioeconomic status

(Marmot et al 1997). In a multivariate analysis, low control in the workplace accounted forabout half the social gradient of coronary heart disease, as adjustment for this factor reduced


29/86

- 29 -

the odds ratio of coronary disease in the low employment group from about 1.4 to about 1.2.

Importantly, the relation between low control and coronary disease was not removed by

adjusting for socioeconomic position (Marmot et al 1997, Bobak et al 1998).

Mediation is important, but it is not the only way in which a variable that predicts disease

incidence in populations can contribute to explaining the social gradient in morbidity and

mortality. The effect modification hypothesis posits that susceptibility to an exposure (such

health-adverse work and employment) is higher among employees in lower socioeconomic

positions compared to higher status people and, therefore, that the effect size produced by the

exposure is higher. The effect modification hypothesis has been tested in several studies

where the effect on health of either high demand and low control at work or of high effort and

low reward at work was found to be greater in lower than in higher socioeconomic groups

(Johnson & Hall 1988, Hallqvist et al 1998, Kuper et al 2002, Wege et al 2008). For instance,

in a German study, depressive symptoms were almost seven times as frequent in the lowest

occupational group scoring high on effort-reward imbalance compared to the highest

occupational group scoring low on effort-reward imbalance (Wege et al 2008).

Effect modification can be observed at different levels of analysis. In a recent comparative

study on welfare regimes, working conditions and health inequalities in different European

countries it was discovered that the effect size of an adverse psychosocial work environment

on health varied according to type of welfare regime. In universalistic welfare states with a

high degree of social protection (Esping-Andersen 1990), effects were smaller compared to

effects in conservative and liberal welfare regimes with less extended social security measures

(Siegrist et al 2009).

In conclusion, evidence in favour of the two hypotheses, mediation and effect modification,

has direct policy implications. Based on the first of these, reducing the adversity of working

conditions and employment could be expected to result in a tangible reduction of the social

gradient across the whole population. The second hypothesis suggests that targetinginterventions towards lower socioeconomic groups, where vulnerability is greatest, would be

expected to reduce the steepness of the social gradient. We conclude that both approaches are

needed.

3.3. Health-promoting effects of work-related interventions

3.3.1. A taxonomy of interventions

Interventions aimed at improving health and well being at work can operate at differentlevels. At the national or supranational level, occupational health and safety legislation and


30/86

- 30 -

distinct policies (e.g. labour market, taxes, education, family and welfare programmes) define

the broader contexts within which more specific actions can be implemented. These specific

interventions can be led by a variety of intermediate organisations, such as employer

associations, trade unions, health services, professional groups concerned with occupational

health and safety, and associations or NGOs dealing with organisational and personnel

development and their business impact. Finally, the intervention needs to be delivered within

either a single organisation or company, a network of interrelated firms or a group of

individuals.

It is common to distinguish primary, secondary and tertiary intervention measures where the

first ones are directed towards general or specific groups of working people who are free from

symptoms of disease or impairment. While secondary interventions address special risk

groups in terms of occupational exposures, behavioural problems or reduced health and

functioning, tertiary interventions deal with rehabilitation and reintegration of (formerly) sick

or chronically ill people or people with longstanding absences from work for other reasons.

At each level of intervention (primary, secondary, tertiary) two distinct, but often combined,

approaches can be taken - work environment change and behavioural change. The latter

applies to groups of people or to individuals.

Interventions that aim to reduce social inequalities in health are commonly located at the level

of primary prevention, and their focus is on the work environment rather than behavioural

change. This is due to the fact that large groups are targeted and that beneficial and cost-

effective interventions might be expected to produce favourable outcomes in a relatively short

time interval. For this reason, the following review gives priority to these types of

intervention. However, special employee assistance or rehabilitation programmes as well as

health-promoting activities at work represent complementary activities with potential impact

on the social gradient of disease (Black 2008).

According to Semmer (2008) work environment change interventions can be classified asenvironment-directed (ergonomic, noise, temperature, work time, broader technological and

organisational context), task-directed (workload, division of work, job autonomy, team work),

and social relationship-directed (communication, conflict, leadership, esteem, social support).

They usually include measures of organisational and personnel development (Noblet &

LaMontangne 2008). In view of the significant contribution of health-adverse behaviours (in

particular cigarette smoking, unhealthy diet and lack of physical activity), in explaining the

social gradient for major chronic diseases (Kivimki et al 2008, Gruer et al 2009), it seemsappropriate to consider the potential for combining health-promoting behavioural change


31/86

- 31 -

programmes and stress management training with approaches that change the work

environment.

In an attempt to systematize the available information on work and employment-related

interventions, where possible we have followed the above structure in putting forward

proposals. That is to say, we start with interventions dealing with employment conditions,

followed by interventions directed towards physical and chemical hazards (including safety

measures). We then address interventions concerned with shift work and other work time

factors, followed by interventions that focus on quality of work in terms of a health promoting

psychosocial work environment.

3.3.2 Employment

Active labour market policies and remuneration and tax policies that aim to reduce income

inequalities are of primary importance in improving the health of the working population. At

the level of national welfare systems, job stability and quality of work were shown to be

greater in universalistic welfare states, such as the Scandinavian welfare system (Dahl et al

2006). Relevant social security arrangements in Scandinavia include setting unemployment

benefits above the poverty threshold. They also include some protection from severe market

forces, an extended pension insurance system based on duration and status of employment,

and relatively generous sickness pay schemes and rehabilitation measures and granting

employees the opportunity to withdraw from work (for short or long periods) because of ill-

health. Workers in routine and manual jobs and employees with lower salaries tend to have a

greater need for such benefits, given their higher workload and poorer general health status

(Dahl et al 2006). Implementing these measures into the English welfare system that is more

strongly characterized by liberal principles (e.g. promotion of private welfare provision,

targeted assistance measures, limited social security) has considerable potential for reducing

health inequalities. The specific areas of intervention are as follows: (a) attempting to reducelong-term unemployment, (b) incentives to increase entry and re-entry into the labour market,

including special programmes for vulnerable groups and (c) investment in the workability and

health of older workers.

For example, a longitudinal study from Sweden found that the health consequences of

unemployment were much smaller among those who were covered by the unemployment

insurance system compared to those without comparable protection (Alm 2001). As full

employment is unrealistic, investment in an informal labour market can be considered as acomplementary strategy for coping with long-term unemployment. Another Swedish study


32/86

- 32 -

documented the beneficial effects on the health of long-term unemployed people of

r

employment and work task group full report

Documents